Sensitive quantitation of endotoxin by enzyme linked immunosorbent assay with monoclonal antibody against Limulus peptide C

Copyright©1994, American Society for Microbiology

Sensitive

Quantitation of Endotoxin by Enzyme-Linked

Immunosorbent

Assay

with Monoclonal Antibody

against Limulus Peptide C

GUI-HANGZHANG,1 LEIF BAEK,2 PETERE. NIELSEN,3 OLE BUCHARDT,4 AND CLAUS

KOCH`*

Department of Immunology, State SerumInstitute, DK-2300CopenhagenS, 1 Departmentof Microbiology,

Herlev Hospital, DK-2730 Herlev,2DepartmentofBiochemistryB, Panum Institute, University of Copenhagen, DK-2200 CopenhagenN,3andDepartmentofChemistry, H. C. 0rstedInstitute,

UniversityofCopenhagen, DK-2100 Copenhagen 0,4 Denmark

Received 17August1993/Returned for modification 22 October 1993/Accepted 10 November1993

Limulus peptide C, a 28-amino-acid fragment of coagulogen formed by the reaction of endotoxin with

Limulus amebocyte lysate,wassynthesized, andamonoclonal antibody against itwasraised. Anewmicroassay

for endotoxin was developed, using this antibody in an enzyme-linked immunosorbent assay for generated peptide C-like

immunoreactivity.

Alinear relationship between absorbance and endotoxin concentrationwas

obtained. Controlstandard endotoxininwatercould be detectedtoalevelof 0.001 endotoxin unitperml. The

endotoxin levels inplasma samples from normalhumans,rabbits, mice, andguinea pigswere

generally

found

tobe below the detection limit of 0.01endotoxinunitpermlof plasma. The color andturbidityof specimens didnotinterfere with theassay.Theconsumption of Limulusamebocytelysatein theassay waslessthan 5%

of that inthegel-clot and chromogenicassays.Withrawlysate, whichwasmuchmorestable in solution than

chloroform-treated lysate, the assay was still highly sensitive to endotoxin butwas totally unresponsive to

natural glucans. The monoclonal antibody cross-reacted with peptide C-like immunoreactivity generated in Tachypleusamebocytelysate,which gaveequal sensitivity in theendotoxinassay.

The Limulus amebocyte lysate (LAL) test is by far the

most sensitive assay forbacterial endotoxins. LAL is

pre-paredfromcirculatingLimulus(horseshoe crab) amebocytes andcontainsacoagulation systemthatmaybeconsidereda

prototype ofmammalian blood coagulation, which involves the sequential activation of several proenzymes (7, 11).

Endotoxin is knowntoactivatethe initialenzyme(factor C)

of the LAL coagulation system, ultimately leading to the conversion ofcoagulogen, aclottableprotein, into coagulin

and peptide C. Visible formation of a coagulin gel-clot

generally indicatesactivation of the LALbyendotoxin and constitutes the basis of the gel-clot method for endotoxin detection. Thegel-clot assayissimpletoperform but lacks

anobjectiveendpointand isnotstrictlyquantitative. Later developments include turbidimetric andchromogenic LAL

assays (8, 24); both are quantitative, objective, and more

sensitive. Kinetic versions of these assays have recently been developed (12, 19), but they are unsuitable for the colored or turbid specimens that are often encountered in clinical and laboratoryuse.

The LALtestwasoriginallyconsideredtobespecificfor endotoxins (11). However, some ,-glucans and

3-glucan-containing mycotic productshavesubsequentlybeenfound to be LAL reactive (1, 2, 4-6, 9). These glucans include curdlan, laminarin,andLAL-reactive materials(2, 16, 17).A recent study indicates that the reactivity of LAL with

1-glucan

is

greatly

influenced

by

the formulation of the LAL

reagent(20).The initiation ofLALcoagulation byendotoxin has been showntobe independentof thatby 3-glucan (13). We have previously developed an endotoxin assay by combining the use of LAL with an enzyme-linked

immu-*Corresponding author. Mailingaddress: Statens Seruminstitut,

Division ofImmunology, 5, Artillerivej, DK-2300Copenhagen S,

Denmark. Phone: +45 3268 3719. Fax: +45 3268 3149.

nosorbent assay (ELISA) for coagulogen (25). This endo-toxinassayissensitive andwell suitedforthedetermination

ofendotoxins inplasma samplesbecause it isnotsubjectto

interferencefrom thecolororturbidityof thespecimens,but ittakesarelatively longtimetoperform (usually5to6h).In thepresentstudy,wehavesynthesizedLimuluspeptideC,a

peptide fragment ofcoagulogen, anddeveloped an ELISA

with amonoclonal antibody (MAb) raisedagainst this

pep-tide. We have found that the generation of peptide C-like immunoreactivityin thecourseof the LAL-endotoxin

reac-tion could be detected much morereadilythan the

conver-sion of coagulogen and correlated well with endotoxin concentrations. Using this assay, we have also tested the reactivities of various LAL preparations to ,B-glucans. We haveconfirmed the functional identityof LAL with Tachy-pleus amebocyte lysate (TAL) for the determination of

endotoxinbymeansof this assay.

MATERIALS ANDMETHODS

All glassware was rendered pyrogen free by heating to

250°C for at least 3 h. Sterile, pyrogen-free tips and

mi-croplateswere purchased fromEppendorf, Hamburg,

Ger-many, and Nunc, Roskilde, Denmark, respectively. LAL reagent waterwas confirmedto have less than 0.001 endo-toxin unit(EU)permlbytheLAL test.

LAL preparations. Commercial LAL preparations in-cludedPyrotell (lot 42-99-541) andPyrotell-T (lot 42-13-575) from Associates ofCapeCod (ACC), Woods Hole, Mass.; LAL (lot 2L0860) from Whittaker Bioproducts,

Walkers-ville, Md.;and LAL(lot 29157-51)from Kabi Vitrum Diag-nostica, Stockholm, Sweden.

Raw amebocyte lysates from Limuluspolyphemus (ob-tained from the MarineBiological Laboratory,WoodsHole, Mass.)andTachypleustridentatus(collectedfrom theBeibu 416

on May 15, 2020 by guest

http://jcm.asm.org/

Gulf of the South China

Sea)

were

prepared

as follows.

Hemolymph

was collected

by

cardiac puncture,

placed

in

pyrogen-free plastic tubes,

and

centrifuged

at

5,000

xgfor 30 minat

4°C.

The cell-free

hemolymph

was

decanted,

and LALreagentwater

equivalent

to

approximately

6volumes of

amebocyte precipitate

was added. The

amebocytes

were

disrupted

by

vigorous

shaking by

hand for 5 min. After

being

kept

at

4°C overnight,

the

lysate

was

centrifuged

at

10,000

x

gfor 30 minat

4°C,

and thesupernatantwasstoredat

-30°C.

Chloroform-treated TAL was

prepared essentially

as de-scribed

previously

(22).

Endotoxin standard. The control standard endotoxin

(CSE)

was

NP-3,

which isa

purified

preparation

of

lipopoly-saccharidefromSalmonella

abortus

equi

(Pyroquant

Diag-nostik, Walldorf, Germany),

and itspotencywasconfirmed

to be 10

EU/ng by comparison

with USP reference standard endotoxin EC-5.

Glucan

preparations.

Laminarin isolated from

Laminania

digitata

wasobtained from

Sigma

(St. Louis, Mo.).

Highly

purified

curdlan,

a

water-insoluble,

carboxymethylated

pow-der,

was

purchased

from Wako Pure Chemical

Industries,

Osaka,

Japan.

Both laminarin andcurdlanweredissolved in 0.2 N NaOH at 5

mg/ml

and incubated at

56°C

for 6 h to

inactivate

contaminating

endotoxin.

MAb

against peptide

C.Limulus

peptide

C, consisting

of 28 amino acid residues

(7,

23),

was

synthesized by

the

solid-phase

"Fmoc" method

(18)

and

conjugated

in amolar ratio of 2:1 to

purified

protein

derivative

(State

Serum

Institute, Copenhagen, Denmark)

with0.1%

glutaraldehyde

(vol/vol)

as the

coupling

reagent. The

conjugate

was ad-sorbed to an aluminum

hydroxide suspension (2 mg/ml

in normal

saline)

and

injected intraperitoneally (0.5

ml per mouse,

equivalent

to 25 p,g of

conjugated peptide)

into female CF1 x

BALB/c

mice. Booster doses were

given

at

2-week intervals. When

high antibody

titerswerefound

by

an ELISAwith

biotinylated

peptide

linkedto avidin-coated

microtiter

plate,

spleen

cells were fused

by

a standard

procedure

(4)

with

myeloma

cells from line

X63/Ag8.6.5.3.

Culture

supernatants

were harvested from

hybridomas

iso-lated after

repeated clonings

by limiting

dilution and screened

by

the ELISAasdescribedabove. Selected MAbs

against peptide

Cwere

purified

from culture

supernatant

by

protein A-Sepharose

CL-4B

affinity

chromatography

(Phar-macia,

Uppsala, Sweden).

Conjugation

ofMAb to horseradish

peroxidase through

a

biotin-avidin

bridge (MAb-baHRP

conjugate).

Onemilliliter of

purified

MAb

(1 mg/ml)

was

dialyzed

against

0.1 M

NaHCO3 overnight

at

4°C

andmixed with 5 ,ulof

N-hydroxy-succinimidobiotin

(40

mg/ml

in

N,N-dimethylformamide)

(Sigma).

After

being

rotatedatroomtemperaturefor 2

h,

the

mixture was

dialyzed

against phosphate-buffered

saline

(PBS)

(pH 7.3) overnight

at

4°C.

The

biotinylated

MAbwas

thenmixedwithan

equal

volume of

streptavidin-horseradish

peroxidase (HRP) conjugate

(Zymed Laboratories,

San

Francisco,

Calif.)

and two volumes of

glycerol

and was

storedat

4°C.

Immunoblotting

of

amebocyte

lysates.

Endotoxin-reacted

amebocyte lysates

were

prepared by incubating

raw

lysates

withan

equal

volume ofCSE

(10 ng/ml)

for1hat37°C.Raw andendotoxin-reacted

amebocyte lysate

samples

werethen

diluted 10-fold in the

sample

buffer

(0.06

M Tris-HCl

[pH

6.8]

containing

10%

glycerol

and2%sodium

dodecyl

sulfate

[SDS])

withorwithout 5%

2-mercaptoethanol

and boiled for 3 min before

being

subjected

to

SDS-polyacrylamide

gel

electrophoresis (SDS-PAGE).

Thiswascarriedoutwith12%

gels, essentially

asdescribed

by

Laemmli

(10),

with

biotiny-lated molecular weight markers (Bio-Rad Laboratories,

Richmond, Calif.). The separated proteins were electroblot-tedontonitrocellulose paper(Schleicher & Schuell,Dassel,

Germany), which was then blocked with PBS containing

0.5% Tween 20 and 0.5 M NaCl. The nitrocellulose paper wasincubated for 1 h with MAb-baHRP conjugate diluted 1:2,000 in PBS containing 0.05% Tween 20 and

streptavidin-HRP(Zymed) diluted 1:4,000 in the same buffer. After the paper was washed with the same buffer, the substrate solution(3,3-diaminobenzidine tetrahydrochloride inbuffer, pH 7.0) (KemEnTec, Copenhagen, Denmark) was added,

and color developmentwas terminated by transferring the nitrocellulose paperto distilled water.

Limuhus peptide C ELISA. Twenty-microliter samples of LAL diluted fourfold in LAL buffer (0.1 M Tris-HCl[pH8.0]

containing 0.15 M NaCl and 0.02 M MgCl2) were addedto the wells ofapyrogen-free microplate and mixed withequal

volumes oftest samples or standards. The plate was then

placed in a hot-plate incubator at 37°C for 20 to 40

min,

depending on the desired sensitivity. The following two types of ELISAwere performed at room temperature after termination of the reaction.

(i) NoncompetitiveELISA. The reaction was stopped by adding 200 ptl of 50 mM NaOH to each well, and 20-,u

aliquotsof the mixturesweretransferredtoamicrotiterplate (Maxisorp; Nunc) towhich 50 mM NaOH (80 ,u per well) hadpreviouslybeen added. After incubation for 30min, the

plate was washed four times with washing buffer (PBS

containing0.5 MNaCl and0.05% TritonX-100, pH 7.3),and 100 pI ofMAb-baHRP

conjugate

at1:2,000indilutionbuffer

(washingbuffercontaining1% bovinealbumin)wasaddedto each well. After incubation for 30min, the platewaswashed fourtimes,and 100pu1of thesubstrate solution (o-phenylene-diaminedihydrochloride [0.4 mg/ml]and0.014% H202 in 0.1 M sodium phosphate-citricacidbuffer, pH 5.0)was added. The colordevelopmentwasstoppedby the addition of 150 pu1

of 1 M

H2SO4,

and the plate was read at 490 nm with a

microplate

reader (Molecular Devices Inc., Menlo Park,

Calif.).

(ii) Competitive

ELISA. Microtiterplateswerecoated (100

pulperwell)with rabbitanti-mouseimmunoglobulins(Z-109;

Dakopatts, Glostrup,

Denmark) diluted 1:2,000 in 10 mM

Tris-HClbuffer(pH 8.6). The LAL-endotoxin reaction was

stopped bythe addition of 100

p,l

ofMAb-baHRP conjugate diluted1:1,000inwashingbuffercontaining10 mM

benzami-dine

(Sigma)

toeachwell.Theplatewasshakenvigorously

for about 5 minon a

shaking

platform, and 100 ,ul of the

contentsof each wellwasthentransferredtothe precoated microtiter

plate,

whichwaswashedfourtimes immediately before use. After incubation for 30 min, the plate was washed four

times,

and the substrate solution (100 plI per

well)

was added. Color development was stopped and the

plate

wasread asdescribed above.

Preparation

and pretreatment ofplasma. Blood from

hu-man and

laboratory

animals was drawn into pyrogen-free

glass

tubes

containing pyrogen-free

heparin (final

concentra-tion,

4

IU/ml).

Plasmawasseparated bycentrifugation at 500

x g for 15 min. Perchloric acid precipitation of plasma to

remove factors that interfere with LAL was carried out

essentially

as described by Obayashi (15) with one minor

modification: the neutralized supernatantwasfurther diluted 1:2 in the LAL reagentwater.

Other LALassays. The

gel-clot

and chromogenic assays

were

performed according

to the manufacturers'

instruc-tions. The

chromogenic

substratefrom the Whittaker

on May 15, 2020 by guest

http://jcm.asm.org/

A

10

97

31

-

_

21

1

1 4

2 3 4 5 6 7 8

B

97 0,1

31*1

21

14 - _

-1 2 3 4 5 6 7 8

FIG. 1. Immunoblotting of LAL (lanes 1 to 4) and TAL (lanes5

to8) before (lanes1and5)and after(lanes2 to 4and6 to8) reaction with endotoxin, detected with the biotinylated MAb against syn-thetic Limulus peptide C. Lanes 2 and 6 contain samples of suspensionsof endotoxin-reacted lysates. Lanes 3 and 7 contain samples of precipitates of the reacted lysates.Lane 4and8contain samplesofsupernatantsof the reactedlysates. The left lane shows biotinylated molecular weight markers; numbers are molecular weights in thousands. (A) Unreduced samples; (B) reduced samples.

mogenic LAL assay kit was used for all the chromogenic assaysin thisstudy.

RESULTS

Immunoblotting of Limulus amebocyte lysates. Figure 1 shows SDS-PAGE immunoblots ofamebocyte lysates be-fore and afterreaction withendotoxin,asdetectedwith the MAb-baHRP conjugate. Without 2-mercaptoethanol

reduc-tion,reacted lysatesandprecipitates ofcentrifugedreacted

lysates gave prominent reactive bands at the 21-kDa posi-tion,while unreactedlysatesand supernatants ofcentrifuged

reactedlysates gavenoreactive band.Withreduction,those reactive bands occurred at the 16-kDa position. The weak bandatthe 21-kDapositionin the reduced unreactedlysates

representscoagulogen,as wasconfirmedby its reaction with MAb against coagulogen, while the prominentband in the endotoxin-reactedsamplesrepresentsaproductwith strong

peptide C-like immunoreactivity.Nobandcorrespondingto free peptide C was observed. Figure 1 also shows the cross-reaction of MAb-baHRP with the peptide C-like im-munoreactivity of reacted TAL.

Standardcurvesof theLimuluspeptide C ELISA.Figure2 shows the standard curves for thetwotypes ofELISA. In

0,01

OD490

0 0.0012 0.0025

0.005

0.01 0.02

Concentration

of CSE (EU/ML)

FIG. 2. Standardcurvesof thenoncompetitiveandcompetitive Limulus peptide C ELISAs, plottedon adoublelogarithmic scale. LAL(ACC;lot42-99-541)wasreacted with CSEfor 40 min at37°C.

the noncompetitive ELISA, in which the generation of

peptideC-likeimmunoreactivitywasmeasureddirectlywith the MAb-baHRP conjugate, the optical density (OD) was

directly proportional to the concentration of endotoxin. In thecompetitive ELISA, inwhich thegenerationofpeptide C-like immunoreactivity was measured indirectly by its inhibition ofbindingof MAb-baHRPconjugatetothewells,

theOD decreased withincreasing endotoxinconcentrations. Similar reaction plots with different sensitivities were ob-tained withcommercialLALsandTALs,andreactionplots

forendotoxinsfrom differentbacterialsources wereparallel (datanotshown).Inthenoncompetitiveassay,totalprotein concentrationsof >0.1mg/mlinthesamplesinterfered with the nonspecific binding of peptide C-like immunoreactivity

to the microtiterwells, leadingtounderestimation of endo-toxin concentrations, whereas no such interference was observed in thecompetitiveassay.Theturbidityorcolor of theoriginal sampledoesnotinterfere with either assay,asit

iswashed outduringtheprocedure.

Theintra-assay coefficient ofvariation, estimatedby

de-terminingtheendotoxincontentofasolution

giving

anOD valueclosetothemidpoint of the standard curve,was10.3%

(n = 6)for thenoncompetitive assayand13.8% (n = 6) for thecompetitive assay.

Effects of reaction time and dilution ofthe LAL and TAL. The sensitivity of the assay with respect to endotoxin increased withincreasing time of theLAL-endotoxin

reac-tion(Fig. 3),anda1:4dilution of LALorTALwasfoundto

givethe best

sensitivity

atthe reaction times shown

(Fig.

4

on May 15, 2020 by guest

http://jcm.asm.org/

OD490

OD490

10-0,1

-0,01

10

0,1

-

0,01-0

0.003

0.012

0.05

0.2

Concentration

of

CSE

(EU/ML)

FIG. 3. Effect of time of reaction of CSE with LAL on the sensitivity of the noncompetitivepeptide C ELISA, plotted on a

double logarithmic scale. The lysate (ACC; lot 42-99-541) was

dilutedin LALbuffer(see MaterialsandMethods) byafactorof5.

and5). With further dilution, an increased susceptibility to

specimen-related interferencewasnoted. The linearportion

of the assaycurveobtained with TAL coveredawider range of endotoxin concentrations than that obtained with LAL. Thesensitivity obtainedat 37°Ccould be achievedat room temperature (220C) by extending the LAL-endotoxin reac-tion timebyabout 30%.

Comparison of sensitivities of different Limulus assays. Table 1comparesthe sensitivities of the Limuluspeptide C ELISA, the gel-clot assay, and thechromogenicassay,using

different commercial LALs and TALs. Thesensitivity ofthe

gel-clotassayisgiven bytheminimum concentration of CSE that produces agel-clot,and that ofthe other assays isgiven bythe minimum concentration of CSE to generate an OD valuesignificantly higherthanthat of the LAL reagent water control. TheLimulus peptide CELISAwas more sensitive

thanthechromogenicorthegel-clotassay.WhenrawLAL orTALwasused,theELISA could detect less than1/50of the minimum concentrations detected by the other assays. Chloroform extraction of theLALincreasedthesensitivities ofthegel-clot and chromogenic assays but was unnecessary fortheLimulus peptide C ELISA.

Stabilityof LAL and TAL in solution. Table 2 shows the effects of chloroform extraction or the addition of dimethyl sulfoxide(DMSO)onthestabilityof LAL insolution at 4°C. Chloroform extraction ofLALhas been used to increase the sensitivities of thegel-clot and chromogenic assays but was

1

*

1:2

0

0.004 0.016 0.063

0.25

1

Concentration of CSE

(EU/ML)

FIG. 4. Effect of dilution of LAL on the sensitivity of the noncompetitivepeptideCELISA,plottedon adoublelogarithmic scale.Thelysate (ACC; lot42-99-541)wasreacted withCSE for40 min at37°C.

shown here to bring about a dramatic reduction of the stabilityof LAL in solution. Addition ofDMSOtothe LAL solution to a concentration of 10% (vol/vol) increased the stability of raw LAL but did not appreciably alter the stabilityofchloroform-treated LAL. Similar effects of chlo-roform and DMSOonthestabilityof TAL in solutionwere also observed.

Endotoxinsin normalplasma.Byusing the PCA methodto pretreat plasma samples, the endotoxin levels in normal

human,mouse,rabbit,andguinea pigplasmaswerefoundto be eitherbeloworclosetothedetection limit(0.01EU/mlof plasma), while those ofpatients with gram-negative sepsis weregenerally above 0.1 EU/ml of plasma(data not shown). Sensitivities of different LAL and TAL preparations to

13-glucans.

Using the Limulus peptide C ELISA,we tested the sensitivities of several commercial and raw LAL and TAL preparations to curdlan, a carboxymethylated

,B-glu-can, andlaminarin, a natural ,-glucan. Different LAL and TAL preparations were classified as chloroform-extracted and nonextracted forms. Table 3 shows that (i)all the LAL and TALpreparations reactedwith curdlan, (ii) the chloro-form-treated preparations of LAL and TAL reacted with bothcurdlan and laminarin,and(iii)therawLALand TAL didnot reactwith laminarin.

When raw TALor LALwas added to commercial chlo-roform-treated LAL (from ACC) in equal amounts, no

changein thereactivityofthe commercialLAL tolaminarin

on May 15, 2020 by guest

http://jcm.asm.org/

TABLE 2. Effects ofchloroform andDMSO treatments on the stability ofLALin solution

Detectionlimit' (EU/ml) Days of

storage Chloroform- Chloroform-treated Raw Raw LAL + at40C treated LAIjb+ 10% LAL 10% DMSO

LAOb DMSO

1 0.004 0.004 0.004 0.004

2 0.004 0.004 0.004 0.004

3 0.008 0.004 0.004 0.004

4 0.032 0.008 0.004 0.004

5 >0.1 0.016 0.004 0.004

10 >0.1 0.004 0.004

30 0.008 0.004

60 0.016 0.004

90 0.032 0.008

aForCSE incubated withLALdiluted 1:4for 30 min (40minforLAL containingDMSO).

b FromACC,lot42-99-541.

of TAL thattherewas a

component

in thelatterthatwas

responsible

for the reactivity. Other chloroform-treated LALs had a

similar effect on the reactivity of raw TAL or LAL to

laminarin.

DISCUSSION

0,01

1:2

_

0 0.004 0.016 0.063 0.025

Concentration

of

CSE

(EU/ML)

FIG. 5. Effect of dilution of TAL on the sensitivity of the noncompetitive peptide C ELISA, plottedon adouble logarithmic scale.The raw TAL(lot 101092)wasreacted withCSEfor30 min at 37°C.

wasobserved,suggestingthat there was no laminarin inhib-itor in the rawpreparations. Interestingly, the raw TAL or LAL becamereactive to50 pg of laminarin per ml when as little as10% ofthecommercialLALwas added, indicating

TABLE 1. Comparison of the sensitivities of different endotoxin assaysbyusing LALorTAL

Detection limit'(EU/mi)with: LAL or TAL source Gel-clot Chromogenic ELISA

assay assay

ACCLALlot42-99-541b 0.06 0.025 0.002 ACC LALlot42-133-575b 0.06 0.031 0.001 KabiVitrumLALlot 0.06 0.012 0.006

29157-51lb

WhittakerBioproducts Notclottable 0.012 0.004 LALlot2LO860C

Raw LALlot051291C 0.50 0.250 0.002

Raw TALlot101092' 0.50 0.250 0.004

Raw TAL lot 150293C 0.50 0.125 0.002

aForCSE under conditionsoptimized for eachassay.

bChloroform-treatedpreparation.

IPreparationwithout chloroformtreatment.

The molecular mechanism ofgel formation in LAL and TAL hasbeenextensively studied by Iwanagaetal. (7, 23). Coagulogen consists ofasingle basicpolypeptidechainwith a calculated molecular mass of 19.7 kDa. It contains three regions, the A chain, peptide C, and the Bchain, of 18, 28, and 129 amino acid residues, respectively. On gelation, peptide C is released and the gel consists oftwochains of A and Bjoined bytwodisulfide bridges. Ourimmunoblotting analyses show that thepeptide-Cimmunoreactivityof

coag-ulogen is formed by the reaction of endotoxin with the LAL. Since there is no apparent loss of molecular weight of coagulogen and nofree peptide C-reactive band, it seems

that peptide C is not split off from coagulogen under our

experimentalconditions. The observed band withpeptide C immunoreactivitymayrepresentan intermediateconsisting

ofpeptide C linkedtotheBchain,assuggested by Tagakiet

al.(23). The intermediate (or peptide C-containing fragment)

TABLE 3. Sensitivities of different LALsorTALstocurdlan and laminarininthenoncompetitivepeptide C ELISA

Detection limit'(ng/ml)

LAL or TAL source

Curdlan Laminarin

ACCLALlot42-99-541b 0.05 0.03

Kabi Vitrum LALlot 0.50 0.50

29157-51lb

TALlot220992b 0.03 0.01

WhittakerBioproducts 1.00 Inactive

LALlot2L0860C

Raw LALlot051291C 0.25 Inactive

RawTAL lot 150293C 0.05 Inactive

a Underconditionsoptimizedfor eachlysate.

bChloroform-treatedpreparation.

cPreparation without chloroformtreatment.

OD490

10-1

-0,1

--+1:8

1:4

*

1:16

on May 15, 2020 by guest

http://jcm.asm.org/

probably occurs in apolymerized form, sinceit is absent in the supernatant and is found only in the insoluble gel of endotoxin-reacted lysates.

Noreaction between MAb-baHRPand unreduced

coagu-logenwasdetected. Reduction of coagulogenwith 2-mercap-toethanol seems to lead to some exposure of the peptide C

epitope, allowing a weak reaction to occur. The use of diluted NaOH in the noncompetitive ELISA serves two

functions, that of terminating the LAL-endotoxin reaction and that ofincreasingthe binding of the peptide C-containing

fragment to the microtiter plate (data not shown). In the

competitive ELISA, benzamidine, a serine protease

inhibi-tor (14), is used tostop theLAL-endotoxin reaction. Thesensitivityofthe Limulus peptide C ELISAwasfound

to depend on the dilution of the LAL reagent and the incubation time of the LAL-endotoxin reaction. The detec-tion limit of theLimulus peptide C ELISA for endotoxin is

mainly restrictedbytheendotoxin content ofLAL. When no limitwassetfor theLAL-endotoxinreactiontime, eightfold-diluted raw LAL gave greater sensitivity than fourfold-diluted LAL, which in turn gave greater sensitivity than twofold-diluted LAL. Oneexplanation of this resultcouldbe that when LALwasdiluted, the endotoxin concentration of the LAL, and hence the assay background, was reduced

accordingly. Another explanation could be that LAL con-tains

endogenous

inhibitorsof the LAL-endotoxin reaction, whose effects arereducedby dilution.

Thestabilityof all LAL assays depends on thestabilityof the LAL reagents. Most commercial LAL reagents have been extracted withchloroform, a method knowntoimprove thesensitivities of thegel-clotand chromogenic assaysby at least afactor of 5 (22). However, chloroform-treated LAL was shown to be much less stable than raw LAL in

solu-tions. Inclusion of 10% DMSO was found to increase the

stability

of raw LAL and TAL in solution. Our results

suggested

that DMSO produced a reversible inhibition of

LALreactivityandaslowinactivation of endotoxin

contam-inating

the LALthattogethermight promote the stabilityof

LAL solutions. No significant differences between endo-toxin concentrations measured with raw LAL or TAL and those measured with DMSO-containing raw LAL or TAL wereobserved, showing that the final DMSO concentration in the reaction mixturewaswithout effect on the endotoxin in thesamples.

The

specificity

of the LAL test has been questioned by

several studies thathave demonstrated reactivity of LALto certain

j-glucans.

Laminarin has been shown to be one of themostpotentLAL-reactive,B-glucans (1). Our results with theLimulus peptideC ELISAconfirmthe finding of

Soder-halletal.

(21),

whoreported reactivity with curdlan but not

with laminarin when an LAL reagent that had not been extracted withchloroformwasused and the contamination of laminarin by endotoxin was eliminated. Roslansky and

Novitsky (20)

reported some reactivity of raw LAL and

non-chloroform-treated LAL with laminarin at concentra-tions of 10to100

,ug/ml.

Ourfindingswere similar, provided that the laminarinwasnotpretreated with NaOH, a proce-duretoinactivateendotoxin.Furthermore, we observed that laminarin totallyblocked the

3-glucan

activation pathway at concentrations of >1 ,ug/ml. These results suggest that the

reactivity

observed by Roslansky and Novitsky (20) was

probably

due to endotoxin contamination of the laminarin

preparation.

The mechanism by which chloroform treatment renders the LAL or TAL reactive to laminarin is not known. Chloroform extraction has been found to increase LAL

sensitivitytoboth endotoxin and 3-glucan, and the removal of lipoproteins that might function as inhibitors has been tentativelyproposed as the underlying mechanism (20,

22).

Since the addition of raw LAL or TAL to commercial chloroform-treated LAL did notaffect the sensitivity ofthe latter to laminarin (unpublished observation), it seems un-likely that there was anylaminarin inhibitor in theraw LAL orTALthat might explain their lackofreactivity.

TAL has been shown to be biochemically similar and functionally identical to LAL (7). In this study, we have demonstrated that the peptide C-containing fragments from LAL andTALare identical inmolecular size and

compara-ble in immunoreactivity. The two lysates show similar

sensitivitiesto endotoxin,similar reactivities tocurdlan and laminarin after chloroform extraction, and no reactivity to laminarin in their raw state. Thus, TAL can be used as a substitute for LAL to detect

endotoxin

both in assays previously described andin the ELISAsdescribed here. The latter present a significant advantage, not only because of their higher sensitivitybut also becausethey reduce LALor TAL consumption to5% of that of the otherassays. This is an important consideration in view of the increasing use of

thesereagentsand thediminishing populationsofrare horse-shoe crab

species

from which

they

are obtained.

ACKNOWLEDGMENTS

Thisstudy was supported by the Danish Biotechnology Centre (a grant to G.-H. Zhang) and the Danish Blood Donor Research Foundation.

We thank Lars Otto Uttenthal for his valuable suggestions during preparation of the manuscript.

REFERENCES

1. Blaschek, W., M. Doll, and G. Franz. 1992. Activation of Limulus amebocyte lysate (LAL) by 3-glucans. Pharm. Phar-macol. Lett. 1:118-122.

2. Carson, L. A., and N. J. Petersen. 1982. LAL-reactive material associated with hemodialysis membranes, p. 217-230. In S. W. Watson, J. Levin, and T. J. Novitsky (ed.), Endotoxins and their detection with the Limulus amebocyte lysate test. Alan R. Liss, Inc., New York.

3. Harlow, E., and D.Lane. 1988. Antibodies: a laboratory man-ual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 4. Hodes, D. S., D. Heon, A. Hass, A. C. Hyatt, and H. L. Hodes. 1987. Reaction of fungal products with amebocytelysates of the Japanese horseshoe crab, Tachypleustridentatus. J. Clin. Mi-crobiol.25:1701-1704.

5. Ikegami, K., K.Ikemura,T. Shimazu, M. Shibuya, H. Sugimoto, T. Yoshioka, and T. Sugimoto. 1988. Early diagnosis of invasive candidiasis and rapid evaluation of antifungal therapy by com-bined use of conventional chromogenic Limulus test and a newly developed endotoxin specific assay. J. Trauma 28:1118-1126.

6. Ikemura, K., K. Ikegami, T. Shimazu, T. Yoshioka, and T. Sugimoto. 1989. False-positive result in Limulus test caused by Limulusamebocytelysate-reactivematerial in immunoglobulin products. J. Clin. Microbiol. 27:1965-1968.

7. Iwanaga, S., T. Miyata, F. Tokunaga, and T. Muta. 1992. Molecular mechanism of hemolymph clotting system in Limu-lus. Thromb. Res. 68:1-32.

8. Iwanaga, S., T. Morita, T. Harada, S. Nakamura, M. Niwa, K. Takada, T. Kimura, and S. Sakakdbara. 1978. Chromogenic substrates for horseshoe crab clotting enzyme. Its application for the assay of bacterial endotoxins. Haemostasis 7:183-188. 9. Kakdnuma, A., T. Asano, H.Toni, and Y. Sugino. 1981. Gelation

of Limulus amebocytelysateby an antitumor

(1,3)-p-D-glucan.

Biochem. Biophys. Res. Commun. 101:434-439.

10. Laemmli, U.K. 1970. Cleavage of structural proteins during the

on May 15, 2020 by guest

http://jcm.asm.org/

assembly of the head ofbacteriophage T4. Nature (London) 227:680-685.

11. Levin, J. 1985. The history of the development of the Limulus amebocytelysatetest, p.3-28. In J. W.Cate, H. R. Buller, A. Sturk, and J. Levin (ed.), Bacterial endotoxin: structure, bio-chemicalsignificance, and detection with the Limulus amebo-cytelysatetest.Alan R. Liss, Inc., New York.

12. Lindsay, G. K., P. F. Roslansky, and T. J. Novitsky. 1989. Single-step, chromogenic Limulus amebocyte lysate assay for endotoxin. J. Clin.Microbiol. 27:947-951.

13. Morita, T., S. Tanaka, T. Nakamura, and S.Iwanaga.1981. A new 1,3-3-D-glucan-mediated coagulation pathway found in Limulusamebocytes. FEBS Lett. 129:318-321.

14. Nakamura, S., T. Morita, T. Harada-Suzuki, S. Iwanaga, K. Takahashi, and M. Niwa. 1982. Aclotting enzyme associated with the hemolymph coagulation system of horseshoe crab (Tachypleus tridentatus): its purification and characterization. J.Biochem. (Tokyo)92:781-792.

15. Obayashi, T. 1984.Addition ofperchloric acidtoblood samples for colorimetric Limulus test using chromogenic substrate: comparison with conventional procedure and clinical applica-tions. J. Lab. Clin. Med. 104:321-330.

16. Pearson, F. C., J. Bohon, W. Lee, G. Bruszer, M.Sagona, G. Jakubowski, R. Dawe, D. Morrison, and C. Dinarello. 1984. Comparison of chemical analyses of hollow-fiber dialyzer ex-tracts.Artif. Organs. 8:291-298.

17. Pearson, F. C., R. Caruana, J. Brukart, D. V. Katz, D. Che-noweth, J.Dubczak, J. Bohon, and M.Weary.1987. The useof theLimulusamebocyte lysate assay to monitor hemodialyzer-associatedsolublecellulosicmaterial (LAL-reativematerial),p. 211-222. In S. W. Watson, J. Levin,and T. J.Novitsky (ed.), Detection ofbacterial endotoxins with the Limulus amebocyte lysatetest.Alan R.Liss, Inc., New York.

18. Plaue,S., and J. P. Briand.1988. Solid-phasepeptidesynthesis, p. 41-94. In M. H. V.VanRegenmortel,J. P.Briand,S.Muller, and S.Plaue(ed.), Synthetic polypeptidesasantigens.Elsevier, Amsterdam.

19. Remillard, J. F., M. C. Gould, P. Roslandsky, and T. J. Novitsky.1987.Quantitationof endotoxininproducts using the LALkineticturbidimetric assay,p.'197-210.InS. W. Watson, J.Levin,andT.J.Novitsky (ed.),Detection ofendotoxinswith theLimulus amebocyte lysate test. Alan R. Liss, Inc., New York.

20. Roslansky,P., and T. J.Novitsky.1991. SensitivityofLimulus amebocyte lysate (LAL) to LAL-reactive glucans. J. Clin. Microbiol. 29:2477-2483.

21. Soderhall,K., J.Levin,and P. B. Armstrong. 1985. Theeffects

of

0-1,3-glucans

onblood coagulation and amebocyte release in

the horseshoe crab,Limuluspolyphemus. Biol. Bull. 169:661-674.

22. Sullivan, J. D., Jr., and S. W. Watson. 1974. Factorsaffecting the sensitivity of Limulus lysate. Appl. Microbiol. 26:1023-1026.

23. Takagi, T., Y. Hokama,S. Iwanaga, S. Nakamura, and M. Niwa. 1979.Amino acid sequencestudiesonhorseshoe crab (Tachy-pleustridentatus) coagulogenandthemechanism of gel forma-tion, p. 169-184.InE.Cohen (ed.), Biomedical applications of the horseshoe crab(Limulidae). AlanR.Liss, Inc.,NewYork. 24. Valois, F. W. 1979. Quantitative method for determining less than apg/ml of LPS,p.415-422. In E.Cohen(ed.), Biomedical applications ofthe horseshoe crab(Limulidae). Alan R. Liss, Inc., New York.

25. Zhang, G.-H., L. Baek, and C. Koch. 1988. Newmicroassay for quantitation ofendotoxinusing Limulus amebocyte lysate

com-binedwithenzyme-linkedimmunosorbent assay. J.Clin. Micro-biol.26:1464-1470.

on May 15, 2020 by guest

http://jcm.asm.org/